AI-Driven Feature Detection and Cost Estimation from 3D CAD Geometries

This dataset contains 1,100 synthetic 3D triangular mesh models for training and evaluating undercut detection using AI models. The dataset includes 600 undercut and 500 non-undercut samples, generated in Blender using parametric scripts as described in the paper "AI-Driven Feature Detection and Cost Estimation from 3D CAD Geometries". All models are provided as OBJ meshes and are suited for research on undercut detection, CAD feature recognition, and manufacturing cost estimation.